John Ryan

VPS 2.0 Outperforms GPS Alone. Niantic Spatial’s Visual Positioning System (VPS) matches live camera data to a spatial model for precise 6DoF localization—indoors, outdoors, and in complex environments where GPS fails. 🔹 ~10-15cm accuracy in mapped areas 🔹 Works in GPS-denied environments 🔹 Persistent spatial anchors Explore this deep dive to learn how VPS works: hubs.ly/Q04csZsd0 #NianticSpatial #Scaniverse #VPS #GIS #GeospatialAI #GEOINT #PhysicalAI #AI #Robotics

Here at [famous scientist’s name].ai, we’re developing tools to accelerate science. Unlike academia, which has stifled the production of high quality scientific work by demanding it be correct, here at [fsn].ai, we know that you can just do (wrong) things.

So we're slowly getting more information on what spooked Google into setting their post-quantum migration timeline to 2029. Two puzzle pieces dropped today.

Incredibly privileged to work with the world class team at Isomorphic on our Drug Design Engine every day. Onward!

A social media platform where you have to do daily reading comprehension tests to be allowed post.

Gemini when asked to analyze current affairs "adapting to fictional reality", may you live in interesting times

I'm very excited for the future we have a gemini flash 3.0 level model (in terms of speed and performance) that is open source and can run on a consumer GPU

I can't believe I need to mute TPU now

I wrote a piece for the @IrishTimes on why Ireland can’t build anything these days, and how to fix that.

We asked every version of Claude to make a clone of Claude(dot)ai, including today’s Sonnet 4.5… see what happened in the video

Maybe to one's surprise, taking KL estimates as `kl_loss` to minimize does *not* enforce the KL. This implementation, however, is quite common in open source RL repos and recent research papers. In short: grad of an unbiased KL estimate is not an unbiased estimate of KL grad.

@cloneofsimo I assume you've seen B.1 in the appendix?

My argument: gist.github.com/cloneofsimo/af… The paper: arxiv.org/abs/2506.05350

Meet MapAnything – a transformer that directly regresses factored metric 3D scene geometry (from images, calibration, poses, or depth) in an end-to-end way. No pipelines, no extra stages. Just 3D geometry & cameras, straight from any type of input, delivering new state-of-the-art results 🚀 One universal model enables SoTA for: 🔥 Mono Depth Estimation 🔥 Multi-View SfM 🔥 Multi-View Stereo 🔥 Depth Completion 🔥 Registration … and many more possibilities! – plus everything is metric 🎯 We release code for data processing, training, benchmarking & ablations – everything Apache 2.0! Details & Links 👇

Flashing custom roms like the old days

Having to shamefully pause until a replacement part arrives after breaking it is the best preparation I had for failed training runs, but both still fill my soul with white hot rage to this day.

I have done the exact same thing, on the same device once again

Scolding Gemini for an incorrect paper implementation from my phone on a 320km/h train with beautiful views. The future is now